bioRxiv Subject Collection: Neuroscience's Journal

Heavy-tailed statistics of cortical representational drift are advantageous for stabilised downstream readouts
Neural representations of familiar environments and mastered tasks continue to change despite no further refinements to task performance or encoding efficiency. Downstream brain regions that depend on a steady supply of information from a neural population subject to this representational drift face a challenge: they must stabilise their readout using only statistical regularities in neural activity. Recent studies have described how representational drift induces deterioration in the accuracy of fixed decoders. Here, we highlight that while a variety of underlying statistics of drift at the level of individual neurons can produce comparable deterioration of a fixed decoder, these different statistics result in dramatically different deterioration rates in adaptive decoders. We describe an adaptive decoding strategy that can read out from a model drifting population long after fixed decoders degrade completely, and demonstrate that in simulated models this strategy is better-suited to heavy-tailed drift statistics, in which individual neurons make sudden and large changes in tuning. We investigate the extent to which these advantageous statistics manifest in experimental in-vivo measurements of drift by considering two existing and well-studied datasets that observe drift in the posterior parietal cortex and the visual cortex. We find preliminary support for sudden jumps in neural tuning that would allow a downstream observer to more easily distinguish changes in representation from noise. These observations are a step towards refining the larger picture of mechanisms underpinning the robustness of information transfer between brain regions that can function in spite of changes in representation driven both by drift and by the learning of new information.

Dysregulation of the fluid homeostasis system by aging
Chronic dehydration is a leading cause of morbidity for the elderly, but how aging alters the fluid homeostasis system is not well understood. Here, we used a combination of physiologic, behavioral and circuit analyses to characterize how fluid balance is affected by aging in mice. We found that old mice have a primary defect in sensing and producing the anti-diuretic hormone vasopressin, which results in chronic dehydration. Recordings and manipulations of the thirst circuitry revealed that old mice retain the ability to sense systemic cues of dehydration but are impaired in detecting presystemic, likely oropharyngeal, cues generated during eating and drinking, resulting in disorganized drinking behavior on short timescales. Surprisingly, old mice had increased drinking and motivation after 24-hour water deprivation, indicating that aging does not result in a general impairment in the thirst circuit. These findings reveal how a homeostatic system undergoes coordinated changes during aging.

Interpreting Sleep Activity Through Neural Contrastive Learning
Memories are spontaneously replayed during sleep, but capturing this process in the human brain has been challenging due to the dominance of slow, rhythmic background activity in sleep, which differs significantly from wakefulness. Each sleep stage, such as NREM and REM, has distinct rhythms, making it even harder for models trained on awake tasks to generalise and decode memory replay during sleep. To overcome this, we developed the Sleep Interpreter (SI), an artificial neural network. We first collected a large EEG dataset from 135 participants, recording brain activity during both awake tasks and overnight sleep. Using a Targeted Memory Reactivation (TMR) technique with 15 pairs of auditory cues and visual images, we tracked when specific memories were reactivated during sleep. The SI model was then trained separately for NREM and REM stages, using contrastive learning to align neural patterns between wakefulness and sleep while filtering out the background rhythms that previously hindered decoding. We also examined how specific sleep rhythms, such as slow oscillations and their coupling with spindles, influenced decoding performance. In a 15-way classification task during sleep, our model achieved a Top-1 accuracy of up to 40.05% on unseen subjects, surpassing all other available decoding models. Finally, we developed a real-time sleep decoding system by integrating an online automatic sleep staging process with the SI model for each sleep stage. This ability to decode brain activity during sleep opens new avenues for exploring the functional roles of sleep. By making our dataset and decoding system publicly available, we provide a valuable resource for advancing research into sleep, memory, and related disorders.

The Simultaneous Measurement of Reversed Phase-Encoding EPI in a Single fMRI Session at 7T: Quantitative and Qualitative Evaluation of Geometric Distortion Correction in Submillimetre fMRI
Motivation: Currently, geometric distortions in EPI are generally corrected using a method that employs the reversed phase-encoding direction. This approach is usually implemented by applying an extra run of the same protocol, only with the phase-encoding direction changed, leading to a substantial increase in redundant acquisition time and specific absorption ratio. Furthermore, while the distortion correction method has been widely employed in numerous fMRI studies, its impact on submillimetre fMRI analysis has remained largely unexplored. Methods: This work presents an EPI scheme that acquires both the original and reversed phase-encoding data in a single fMRI session. The feasibility of using the method for submillimetre EPI (0.73 x 0.73 mm2) was verified with visual fMRI at 7T. EPI distortions were corrected using the ANTs software, and its performance was evaluated using various criteria, including spatial resolution, functional mapping accuracy, and histogram distribution. Results: The presented scheme effectively reduced redundant acquisition time and therefore total radio-frequency energy. The distortion-corrected fMRI data demonstrated significant improvements in co-registration with anatomical scans and functional mapping accuracy. As a result, the ratio of grey-matter-activated voxels was substantially enhanced (on average, 88.58%). These improvements were achieved without significant degradation of spatial resolution or alternation of the functional activation distribution; a high degree of similarity between the original and distortion-corrected cases ({rho} > 0.99) was observed in the t-value histograms. Conclusions: This work presents a simultaneous acquisition scheme for reversed phase-encoding EPI, demonstrating its effectiveness and the benefits achieved through distortion correction in submillimetre fMRI at 7T.

Hypothalamic connectivities and self-evaluated aggression in young adults
Introduction: The hypothalamus plays a pivotal role in supporting motivated behavior, including aggression. Previous work suggested differential roles of the medial hypothalamus (MH) and lateral hypothalamus (LH) in aggressive behaviors, but little is known about how their resting-state functional connectivity (rsFC) may relate to aggression in humans. Methods: We employed the data from the Human Connectome Project (HCP) and examined the rsFC's of LH and MH in 745 young adults (393 women). We also explored sex differences in the rsFC's. We processed the imaging data with published routines and evaluated the results of voxel-wise regression on aggression score, as obtained from Achenbach Adult Self Report, with a corrected threshold. Results: The analysis revealed significant rsFC between the LH and clusters in the middle temporal and occipital gyri across all subjects and in the thalamus for men, both in negative correlation with aggression score. Slope test confirmed sex differences in the correlation between LH-thalamus rsFC and aggression score. No significant rsFC was observed for MH. Conclusions: These findings suggest a role of LH rsFCs and sex differences in LH-thalamus rsFC in the manifestation of aggression in humans. The findings highlight the need for further research into sex-specific neural pathways in aggression and other related behavioral traits of importance to mental health.

A Unified Theory of Response Sparsity and Variability for Energy-Efficient Neural Coding
Understanding how cortical neurons use dynamic firing patterns to represent sensory signals is a central challenge in neuroscience. Decades of research have shown that cortical neuronal activities exhibit high variance, typically quantified by the coefficient of variation (CV), suggesting intrinsic randomness. Conversely, substantial evidence indicates that cortical neurons display high response sparseness, indicative of efficient encoding. The apparent contradiction between these neural coding properties, stochastic yet efficient, has lacked a unified theoretical framework. This study aims to resolve this discrepancy. We conducted a series of analyses to establish a direct relational function between CV and sparseness, proving they are intrinsically correlated or equivalent across different statistical distributions in neural activities. We further derive a function showing that both irregularity and sparsity in neuronal activities are positive functions of energy-efficient coding capacity, quantified by Information-Cost Efficiency (ICE). This suggests that the observed high irregularity and sparsity in cortical activities result from a shared mechanism optimized for maximizing information encoding capacity while minimizing cost. Furthermore, we introduce a CV-maximization algorithm to generate kernel functions replicating the receptive fields of the primary visual cortex. This finding indicates that the neuronal functions in the visual cortex are optimal energy-efficient coding operators for natural images. Hence, this framework unifies the concepts of irregularity and sparsity in neuronal activities by linking them to a common mechanism of coding efficiency, offering deeper insights into neural coding strategies.

Adult choline supplementation in a Down syndrome model reduces co-morbidities and improves cognition
Down syndrome (DS) is the most common cause of early-onset Alzheimers disease (AD). Dietary choline has been proposed as a modifiable factor to improve cognitive and pathological outcomes of AD, especially as many do not reach adequate daily intake levels. Perinatal choline supplementation (Ch+) in the Ts65Dn mouse model of DS protects offspring against AD-relevant pathology and improves cognition, and dietary Ch+ in adult AD models also ameliorates pathology and improves cognition. However, dietary Ch+ in adult Ts65Dn mice has not yet been explored. To test whether Ch+ in adulthood improves DS co-morbidities, we fed trisomic Ts65Dn mice and disomic littermate controls with either choline normal (ChN; 1.1 g/kg) or Ch+ (5 g/kg) diets from 4.5-14 months of age. We found that Ch+ improved cognitive flexibility in a reverse place preference task and reduced weight gain and peripheral inflammation in female mice, whereas Ch+ improved glucose metabolism in male mice. In conclusion, we found that adulthood Ch+ benefits behavioral and biological factors important for general well-being in DS and related to AD risk.

Executive dysfunction is associated with altered hippocampal-prefrontal functional connectivity in male 3xTg Alzheimer's model mice
Executive function depends on connectivity between the ventral hippocampus and medial prefrontal cortex (mPFC). How abnormalities in this pathway lead to cognitive dysfunction in Alzheimer's disease (AD) have yet to be elucidated. Here, male 3xTg AD mice at 6-months displayed maladaptive decision-making in the rodent 4-Choice Gambling Task measure of executive function. Extracellular field recordings in the infralimbic cortex at this age showed layer-specific reductions in response amplitude and paired-pulse ratio following activation of hippocampal input fibres, indicating changes to short-term hippocampal-prefrontal synaptic plasticity. Bulk RNA sequencing of the mPFC in 6-month-old mice identified differential gene expression changes associated with calcium ion transport, glutamatergic, GABAergic, and dopaminergic neurotransmission. Seven of these genes (Gpm6b, Slc38a5, Ccr5, Kcnj10, Ddah1, Gad1, Slc17a8) were also differentially expressed in 3-month mice. These results reveal a pre-clinical deficit in executive function correlating with synaptic plasticity and gene expression changes in the mPFC of male 3xTg mice.

Adaptation of retinal discriminability to natural scenes
Sensory systems discriminate stimuli to direct behavioral choices, a process governed by two distinct properties - neural sensitivity to specific stimuli, and stochastic properties that importantly include neural correlations. Two questions that have received extensive investigation and debate are whether visual systems are optimized for natural scenes, and whether noise correlations contribute to this optimization. However, the lack of sufficient computational models has made these questions inaccessible in the context of the normal function of the visual system, which is to discriminate between natural stimuli. Here we take a direct approach to analyze discriminability under natural scenes for a population of salamander retinal ganglion cells using a model of the retinal neural code that captures both sensitivity and stochasticity. Using methods of information geometry and generative machine learning, we analyzed the manifolds of natural stimuli and neural responses, finding that discriminability in the ganglion cell population adapts to enhance information transmission about natural scenes, in particular about localized motion. Contrary to previous proposals, noise correlations reduce information transmission and arise simply as a natural consequence of the shared circuitry that generates changing spatiotemporal visual sensitivity. These results address a long-standing debate as to the role of retinal correlations in the encoding of natural stimuli and reveal how the highly nonlinear receptive fields of the retina adapt dynamically to increase information transmission under natural scenes by performing the important ethological function of local motion discrimination.

Effect of chronic upregulation of endocannabinoid signaling in vivo with JZL184 on striatal synaptic plasticity and motor learning in YAC128 Huntington disease mice
Synaptic dysfunction underlies early sensorimotor and cognitive deficits, and precedes neurodegeneration in a variety of disorders, including Alzheimer, Parkinson and Huntington disease (HD). A monogenic inherited disorder, HD manifests with cognitive, motor and mood disorders associated with progressive degeneration of striatal spiny projection neurons and cortical pyramidal neurons. Cortico-basal ganglia-thalamic loops regulate movement selection and motor learning, which are impaired early in HD. Skilled motor learning is mediated in part by plasticity at cortico-striatal synapses, including endocannabinoid-mediated, high-frequency stimulation induced long-term depression (HFS-LTD). Previously, we found impaired HFS-LTD in brain slice recordings from pre-manifest HD mouse models, which was corrected by JZL184, an inhibitor of endocannabinoid 2-arachidonoyl glycerol (2-AG) degradation. Here, we tested the effects of JZL184 administered in vivo to YAC128 HD model and wild-type (WT) littermate mice. JZL184, given orally daily over a 3-week period, significantly increased levels of 2-AG in striatal tissue. While JZL184 treatment had no impact on open field behavior which was similar for the two genotypes, the treatment improved motor learning on the rotarod task in YAC128 mice to the level observed in WT mice. Moreover, HFS-induced striatal plasticity measured by field potential recording in acute brain slice from YAC128 mice was normalized to WT levels after JZL184 treatment. These results suggest a novel target for mitigating early symptoms of HD, and support the need for clinical trials to test the efficacy of modulating the endocannabinoid system in treatment of HD.

Targeted single cell expression profiling identifies integrators of sleep and metabolic state
Animals modulate sleep in accordance with their internal and external environments. Metabolic cues are particularly potent regulators of sleep, allowing animals to alter their sleep timing and amount depending on food availability and foraging duration. The fruit fly, Drosophila melanogaster, suppresses sleep in response to acute food deprivation, presumably to forage for food. This process is dependent on a single pair of Lateral Horn Leucokinin (LHLK) neurons, that secrete the neuropeptide Leucokinin. These neurons signal to insulin producing cells and suppress sleep under periods of starvation. The identification of individual neurons that modulate sleep-metabolism interactions provides the opportunity to examine the cellular changes associated with sleep modulation. Here, we use single-cell sequencing of LHLK neurons to examine the transcriptional responses to starvation. We validate that a Patch-seq approach selectively isolates RNA from individual LHLK neurons. Single-cell CEL-Seq comparisons of LHLK neurons between fed and 24-hr starved flies identified 24 genes that are differentially expressed in accordance with starvation state. In total, 12 upregulated genes and 12 downregulated genes were identified. Gene-ontology analysis showed an enrichment for Attacins, a family of anti-microbial peptides, along with a number of transcripts with diverse roles in regulating cellular function. Targeted knockdown of differentially expressed genes identified multiple genes that function within LHLK neurons to regulate sleep-metabolism interactions. Functionally validated genes include an essential role for the E3 ubiquitin ligase insomniac, the sorbitol dehydrogenase Sodh1, as well as AttacinC and AttacinB in starvation-induced sleep suppression. Taken together, these findings provide a pipeline for identifying novel regulators of sleep-metabolism interactions within individual neurons.

Evaluation of HD-sEMG descriptor sensitivity to changes of anatomical and neural properties with aging : A simulation study
Background and objectives: A reliable evaluation of anatomical and neural muscle properties and its effects on the electrical signals measured at the skin surface aims to develop a medical non-invasive aid-diagnosis tool assisted by model and personalized to the patient. This tool will be dedicated to understand and evaluate muscle diseases and aging. Methods: We perform a new Robust Morris Screening Method: RMSM in Douania et al. (2023), to assess the impact of muscle anatomy (model inputs) uncertainties and variations on a simulated HD-sEMG signals (model outputs). The model describes a complex neuromuscular system simulating HD-sEMG (high density surface electromyography) signals generated from motor units electrical sources of a striated muscle: the Biceps Brachii (BB). Two subjects categories and two contractions levels are studied: young men (YM) and old men (OM) at low and high contraction (LC = 20% of MVC and HC = 60% of MVC). A 33 features in time and frequency domains are used as model outputs. Results: We have demonstrated that the neuromuscular model is able to deliver HD-sEMG signals sensitive to the same anatomical and neural muscle factors as in real cases. Time domain features are mainly sensitive to muscle thickness, conduction velocity of fibers, electrode locations (at HC), number of motor units, and to the number of slow and fast fibers for young and aged categories respectively. Frequency domain feature are sensitive mainly to the conduction velocity of fibers and muscle conductivities (no significant differences between YM and OM are observed). Conclusion: This result is important, it allows to obtain simulated HD-sEMG signals close to experimental ones with low cost and in reduced time. However, for a reliable evaluation of muscle aging, the neuromuscular model should be enhanced to better describe structural, morphological, and functional age-related phenomena.

Coordinated multi-level adaptations across neocortical areas during task learning
The coordinated changes of neural activity during learning, from single neurons to populations of neurons and their interactions across brain areas, remain poorly understood. To reveal specific learning-related changes, we applied multi-area two-photon calcium imaging in mouse neocortex during training of a sensory discrimination task. We uncovered coordinated adaptations in primary somatosensory area S1 and the anterior (A) and rostrolateral (RL) areas of posterior parietal cortex (PPC). At the single-neuron level, task-learning was marked by increased number and stabilized responses of task neurons. At the population level, responses exhibited increased dimensionality and reduced trial-to-trial variability, paralleled by enhanced encoding of task information. The PPC areas, especially area A, became gradually engaged, opening additional within-area subspaces and inter-area subspaces with S1. Task encoding subspaces gradually aligned with these interaction subspaces. Behavioral errors correlated with reduced neuronal responses, decreased encoding accuracy, and misaligned subspaces. Thus, multi-level adaptations within and across cortical areas contribute to learning-related refinement of sensory processing and decision-making.

Sleep modulates neural timescales and spatiotemporal integration in the human cortex
Spontaneous neural dynamics manifest across multiple timescales, which are intrinsic to brain areas and exhibit hierarchical organization across the cortex. In wake, a hierarchy of timescales is thought to naturally emerge from microstructural properties, gene expression, and recurrent connections. A fundamental question is timescales' organization and changes in sleep, where physiological needs are different. Here, we describe two coexisting but distinct measures of neural timescales, obtained from broadband activity and gamma power, which display complementary properties. We leveraged intracranial electroencephalography (iEEG) data to characterize timescale changes from wake to sleep across the cortical hierarchy. We show that both broadband and gamma timescales are globally longer in sleep than in wake. While broadband timescales increase along the sensorimotor-association axis, gamma ones decrease. During sleep, slow waves can explain the increase of broadband and gamma timescales, but only broadband ones show a positive association with slow-wave density across the cortex. Finally, we characterize spatial correlations and their relationship with timescales as a proxy for spatiotemporal integration, finding high integration at long distances in wake for broadband and at short distances in sleep for gamma timescales. Our results suggest that mesoscopic neural populations possess different timescales that are shaped by anatomy and are modulated by the sleep/wake cycle.

Undoing of firing rate adaptation enables invariant population codes
Adaptation often implements an efficient coding strategy to align neuron sensitivity to relevant stimuli. However, when sensory information is encoded in populations of neurons, adaptation of individual units could deteriorate stimulus information relevant for behavior. One prominent example is the encoding of olfactory information, which occurs at the population level. We show that if individual ORNs were to adapt following an efficient coding principle, contrast information would be lost at the population level, impairing the detection of ON and OFF stimuli. Surprisingly, adaptation of ORN firing responses is compensated at the axon terminal, where calcium transients are kept background-invariant by inhibitory feedback. We demonstrate that achieving this invariance across different background conditions requires an adaptation strategy that shifts response amplitude, rather than sensitivity as predicted by the efficient coding principle. This background invariance is passed on to second-order olfactory neurons, through facilitation of vesicle release that involves modulation of Unc13 proteins. We conclude that synaptic and circuit computations compensate peripheral firing rate adaptation, enhancing the separation of ON and OFF contrasts, while preserving the identity of ON stimuli in population codes. Our findings identify a strategy that allows neural circuits to minimize the cost of information transmission while preserving information content.

Minimizing the number of phosphenes required for object recognition under prosthetic vision
Cortical prostheses offer the potential for partial vision restoration in individuals with blindness by stimulating V1 neurons to produce phosphenes. However, the low number of phosphenes that can be elicited in practice makes encoding of whole objects difficult, and the round shape of phosphenes lack the contour cues necessary for perceptual grouping. We propose a minimalistic encoding approach that focuses on essential visual information. We fragmented objects' contours into either phosphenes or curved segments, providing either low or high local visual information. 46 participants identified these fragmented objects in a free-naming task. The number of fragments gradually increased to quantify the minimum number of phosphenes and segments necessary to recognize objects. Most objects could be recognized with only 65 phosphenes, which is in the range of implantable electrodes in human patients. Participants required 27% fewer segments than phosphenes to recognize objects. Including individual objects as a random effect in a linear mixed model substantially increased the explained variance, suggesting that the minimal number of fragments required for object recognition in prosthetic vision strongly depends on the particular object. Our results demonstrate that a minimalistic approach can substantially reduce the number of phosphenes required for recognition, emphasizing the importance of identifying critical object features to minimize brain stimulation in visual prostheses.

The neural time course of size constancy in natural scenes
Accurate real-world size perception relies on size constancy, a mechanism that integrates an object's retinal size with distance information. The neural time course of extracting pictorial distance cues from scenes and integrating them with retinal size information - a process referred to as scene-based size constancy - remains unknown. In two experiments, participants viewed objects with either large or small retinal sizes, presented at near or far distances in outdoor scene photographs, while performing an unrelated one-back task. We applied multivariate pattern analysis (MVPA) to time-resolved EEG data to decode the retinal size of large versus small objects, depending on their distance (near versus far) in the scenes. The objects were either perceptually similar in size (large-near versus small-far) or perceptually dissimilar in size (large-far versus small-near), reflecting size constancy. We found that the retinal size of objects could be decoded from 80 ms after scene onset onwards. Distance information modulated size decoding at least 120 ms later: from 200 ms after scene onset when objects were fixated, and from 280 ms when objects were viewed in the periphery. These findings reveal the neural time course of size constancy based on pictorial distance cues in natural scenes.

Altered mRNA transport and local translation in iNeurons with RNA binding protein knockdown
Neurons rely on mRNA transport and local translation to facilitate rapid protein synthesis in processes far from the cell body. These processes allow precise spatial and temporal control of translation and are mediated by RNA binding proteins (RBPs), including those known to be associated with neurodegenerative diseases. Here, we use proteomics, transcriptomics, and microscopy to investigate the impact of RBP knockdown on mRNA transport and local translation in iPSC-derived neurons. We find thousands of transcripts enriched in neurites and that many of these transcripts are locally translated, possibly due to the shorter length of transcripts in neurites. Loss of frontotemporal dementia/amyotrophic lateral sclerosis (FTD/ALS)-associated RBPs TDP-43 and hnRNPA1 lead to distinct alterations in the neuritic proteome and transcriptome. TDP-43 knockdown (KD) leads to increased neuritic mRNA and translation. In contrast, hnRNPA1 leads to increased neuritic mRNA, but not translation, and more moderate effects on local mRNA profiles, possibly due to compensation by hnRNPA3. These results highlight the crucial role of FTD/ALS-associated RBPs in mRNA transport and local translation in neurons and the importance of these processes in neuron health and disease.

Novel color vision assessment tool: AIM Color Detection and Discrimination
Color vision assessment is essential in clinical practice, yet different tests exhibit distinct strengths and limitations. Here we apply a psychophysical paradigm, Angular Indication Measurement (AIM) for color detection and discrimination. AIM is designed to address some of the shortcomings of existing tests, such as prolonged testing time, limited accuracy and sensitivity, and the necessity for clinician oversight. AIM presents adaptively generated charts, each a N*M (here 4*4) grid of stimuli, and participants are instructed to indicate either the orientation of the gap in a cone-isolating Landolt C optotype or the orientation of the edge between two colors in an equiluminant color space. The contrasts or color differences of the stimuli are adaptively selected for each chart based on performance of prior AIM charts. In a group of 23 color-normal and 15 people with color vision deficiency (CVD), we validate AIM color against Hardy-Rand-Rittler (HRR), Farnsworth-Munsell 100 hue test (FM100), and anomaloscope color matching diagnosis and use machine learning techniques to classify the type and severity of CVD. The results show that AIM has classification accuracies comparable to that of the anomaloscope, and while HRR and FM100 are less accurate than AIM and an anomaloscope, HRR is very rapid. We conclude that AIM is a computer-based, self-administered, response-adaptive and rapid tool with high test-retest repeatability that has the potential to be suitable for both clinical and research applications.

New neural-inspired controller generalises modularity over lower limb tasks through internal models
Predictive neuromuscular models based on neural controllers are a powerful tool for testing assumptions on the underlying architecture of sensorimotor control and its associated neural activity. However, most current controllers suffer from lack of physiological plausibility and are generally task specific. We propose a new neural controller, called Internal Model-based Modular Controller (IMMC), where a hierarchical architecture organises generalizable modules in activation networks dedicated to different motion tasks. The architecture comprises a simple model of the mesencephalic locomotor region (MLR), which sends controlling signals that manage the activity of internal models (IMs). The IMs organise synergies, coordinated and stereotyped activity of multiple muscles, in task-specific networks. The resultant organisations allow the generalisation of this architecture to different lower limb motions. The IMMC was tested in Stand-To-Walk simulations (STW), where the MLR switches between two IMs that recombine five synergies to replicate the standing and walking tasks. The simulation kinematics, muscle activation patterns and ground reaction forces were generally consistent with experimental data. In addition, the controller can transition to slower and faster speeds by tuning a single controlling signal. The proposed architecture is a first step to develop neuromuscular models which integrate multiple motor behaviours in a unified controller.

The claustrum enhances neural variability by modulating the responsiveness of the prefrontal cortex
The claustrum is recognized for its significant impact on various cognitive functions and its extensive connections with other brain regions, yet its functional role remains to be fully understood. Here, we utilized an optogenetic approach to investigate the claustrum's influence on neuronal activity within the dorsal prefrontal cortex (dPFC) of mice. We conducted two-photon calcium imaging to assess dPFC cell responses during exposure to visual stimuli and widefield photostimulation of claustrum axons embedded in the dPFC. We identified three distinct subpopulations of neurons - sensory responsive, opto responsive, and opto-boosted cells - each exhibiting unique response dynamics to combined visual and optogenetic stimuli. Our findings reveal that stimulation of claustrum axons can normalize neuronal responsiveness, while enhancing neural variability, and significantly increasing network homogeneity. Training in a Pavlovian task showed that while enhanced variability with claustrum axon stimulation in neural responses persists, training does not further increase this variability but instead leads to greater network homogeneity. Additionally, we also performed claustrum axon silencing experiments that revealed that the claustrum may operate bidirectionally to maintain enhanced variability and homogeneity in the prefrontal cortex. These results highlight the crucial role of the claustrum in dynamically modulating dPFC activity, impacting both neuronal variability and network synchronization.

Millisecond-scale motor control precedes sensorimotor learning in Bengalese finches
A key goal of the nervous system in young animals is to learn motor skills. Songbirds learn to sing as juveniles, providing a unique opportunity to identify the neural correlates of skill acquisition. Prior studies have shown that spike rate variability decreases during song acquisition, suggesting a transition from rate-based neural control to the millisecond-precise motor codes known to underlie adult vocal performance. By quantifying how the ensemble of spike patterns fired by cortical neurons (the ``neural vocabulary'') and the relationship between spike patterns and song acoustics (the ``neural code'') change during song acquisition, we quantified how vocal control changes across learning in juvenile Bengalese finches. We found that despite the expected drop in rate variability (a learning-related change in spike vocabulary), the precision of the neural code in the youngest singers is the same as in adults, with 1--2 millisecond variations in spike timing transduced into quantifiably different behaviors. In contrast, fluctuations of firing rates on longer timescales fail to affect the motor output. The consistent presence of millisecond-scale motor coding during changing levels of spike rate and behavioral variability supports the view that variability early in learning stems from deliberate motor exploration rather than imprecise motor control.

Activation of group I mGluRs is required for heterosynaptic priming of long-term potentiation in mouse hippocampus
Metaplasticity involves changes in the state of neurons or synapses that influence their ability to generate synaptic plasticity. One form of heterosynaptic metaplasticity, known as synaptic tagging and capture (STC), has been intensively studied but the underlying mechanisms are not fully understood. In experiments using hippocampal slices prepared from C57BL/6J mice, we have examined the role of group I metabotropic glutamate receptors (mGluRs) in STC. We used a version of STC where a strong theta-burst stimulus (TBS), delivered to one set of Schaffer collateral-commissural pathway inputs to CA1, preceded a weak TBS delivered to a second independent set of inputs. We observed that, firstly, dual inhibition of mGluR1 and mGluR5, using YM 298198 and MTEP respectively, did not affect a form of protein synthesis-independent LTP (LTP1), but substantially inhibited a form of protein synthesis-dependent LTP (LTP2). Secondly, these inhibitors prevented the small heterosynaptic potentiation, which is often associated with LTP2. Thirdly, STC was abolished when these antagonists were applied either during the strong (priming) TBS or during the subsequent weak TBS at the independent pathway. It is proposed that the activation of group I mGluRs serves as a trigger for local protein synthesis both during the strong and weak TBS and, as such, are an integral part of the STC process, a process involved in associative learning and memory.

Purkinje cell intrinsic activity shapes cerebellar development and function
The emergence of functional cerebellar circuits is heavily influenced by activity-dependent processes. However, the role of intrinsic activity in Purkinje neurons, independent of external input, in driving cerebellar development remains less understood. Here, we demonstrate that before synaptic networks mature, Purkinje cell intrinsic activity is essential for regulating dendrite growth, establishing connections with cerebellar nuclei, and ensuring proper cerebellar function. Disrupting this activity during the postnatal period impairs motor function, with earlier disruptions causing more severe effects. Importantly, only disruptions during early development lead to pronounced defects in cellular morphology, highlighting key temporal windows for dendritic growth and maturation. Transcriptomic analysis revealed that early intrinsic activity drives the expression of activity-dependent genes, such as Prkc{gamma} and Car8, which are essential for dendritic growth. Our findings emphasize the importance of temporally-specific intrinsic activity in Purkinje cells for guiding cerebellar circuit development, providing a potential common mechanism underlying cerebellum-related disorders.

Impaired excitability of fast-spiking neurons in a novel mouse model of KCNC1 epileptic encephalopathy
The recurrent pathogenic variant KCNC1-p.Ala421Val (A421V) is a cause of developmental and epileptic encephalopathy characterized by moderate-to-severe developmental delay/intellectual disability, and infantile-onset treatment-resistant epilepsy with multiple seizure types including myoclonic seizures. Yet, the mechanistic basis of disease is unclear. KCNC1 encodes Kv3.1, a voltage-gated potassium channel subunit that is highly and selectively expressed in neurons capable of generating action potentials at high frequency, including parvalbumin-positive fast-spiking GABAergic inhibitory interneurons in cerebral cortex (PV-INs) known to be important for cognitive function and plasticity as well as control of network excitation to prevent seizures. In this study, we generate a novel transgenic mouse model with conditional expression of the Ala421Val pathogenic missense variant (Kcnc1-A421V/+ mice) to explore the physiological mechanisms of KCNC1 developmental and epileptic encephalopathy. Our results indicate that global heterozygous expression of the A421V variant leads to epilepsy and premature lethality. We observe decreased PV-IN cell surface expression of Kv3.1 via immunohistochemistry, decreased voltage-gated potassium current density in PV-INs using outside-out nucleated macropatch recordings in brain slice, and profound impairments in the intrinsic excitability of cerebral cortex PV-INs but not excitatory neurons in current-clamp electrophysiology. In vivo two-photon calcium imaging revealed hypersynchronous discharges correlated with brief paroxysmal movements, subsequently shown to be myoclonic seizures on electroencephalography. We found alterations in PV-IN-mediated inhibitory neurotransmission in young adult but not juvenile Kcnc1-A421V/+ mice relative to wild-type controls. Together, these results establish the impact of the recurrent Kv3.1-A421V variant on neuronal excitability and synaptic physiology across development to drive network dysfunction underlying KCNC1 epileptic encephalopathy.

The trade-off between temporal precision and effect amplitude of inhibitory plasticity regulation determines separability of learned representations
Synaptic plasticity, the process by which synapses change in an activity-dependent manner, is assumed to be the basis of learning. Experimental evidence demonstrates that activity originating from other synapses in close proximity to an observed one can influence the outcome of plasticity including activity from inhibitory synapses. Under the assumption that the regulatory effect of inhibition is mediated by hyperpolarisation, we identify a trade-off between temporal precision and effect amplitude due to the treatment of postsynaptic activity in three different voltage-dependent plasticity models. Generally, inhibitory regulation of plasticity enhances the competition between lateral neurons driving the development of functionally relevant connectivity structures in recurrent excitatory-inhibitory networks. Thus, all models show signs of the ability to perform Independent Component Analysis (ICA) and lead to receptive field development. Models which are highly sensitive to local synaptic information tend to result in a higher degree of separation between learned features. This work stresses the importance of considering inhibition in plasticity research as well as indicates that learned representations are influenced by the available information at a synaptic site.

Positive and Negative Retinotopic Codes in the Human Hippocampus
The hippocampus is thought to coordinate sensory-mnemonic information streams in the brain, representing both the apex of the visual processing hierarchy and also a central hub of mnemonic processing. Yet, the mechanisms underlying sensory-mnemonic interactions in the hippocampus are poorly understood. Recent work in cortex suggests that a retinotopic code - typically thought to be exclusive to visual areas - may help organize internal and external information at the cortical apex via opponent interactions. Here, we leverage high-resolution 7T functional MRI to test whether a bivalent retinotopic code structures activity within the human hippocampus, and mediates hippocampal-cortical interactions. Our findings reveal a robust retinotopic code in the hippocampus, characterized by stable population receptive fields (pRFs) with consistent preferred visual field locations across experimental runs. Notably, this retinotopic code comprises roughly equal proportions of positive and negative pRFs, aligning with the hypothesized role of negative pRFs in mnemonic processing. Finally, the signed amplitude of hippocampal pRFs predicts functional connectivity between retinotopic hippocampal and cortical voxels. Taken together, these results suggest that retinotopic coding may scaffold internal mnemonic and external sensory information processing within the hippocampus, and across hippocampal-cortical interactions.

The Neural Blueprint of Stress Susceptibility: Brain-wide neuronal activity associated with the consequences of stress
Understanding the neurobiological mechanisms of stress susceptibility is key to advancing our insight into stress-related psychopathology like post-traumatic stress disorder (PTSD). Preclinical animal models however typically lack translationally relevant brain readouts. Here, we used a mouse model for severe stress, segregating mice into stress-susceptible and resilient groups. We analyzed and contrasted their whole-brain neuronal activity pre-, peri- and post-stress exposure and compared functional connectivity of the salience (SN), default mode (DMN) and executive control network (lateral cortical network (LCN) in rodents). We found that stress-susceptible mice exhibited pre-existing hyperactivity in the lateral orbital area, a potential risk factor for stress vulnerability, and heightened retrosplenial cortex activity peri- and post-stress, potentially contributing to maladaptive fear memory. Upon stress exposure, susceptible mice showed strong recruitment of visual and memory-related areas, whereas resilient mice displayed marked reductions in retrosplenial activity and increased activation in the agranular insula and ventral striatum. Observations of enhanced intra-SN and SN-DMN connectivity in susceptible mice mirrored those observed in individuals with PTSD. Moreover, susceptible mice displayed aberrant DMN-LCN network pre- and peri-stress. These findings highlight the importance of dynamic network interactions in stress susceptibility and suggest novel brain region targets for (early) intervention.

Dynamic cognitive differences between internal and external attention are associated with depressive and anxiety disorders
The internal/external attention framework characterises attention focused on internal representations, such as emotions, versus external representations, such as perceptual stimuli. The inability to focus one's attention is considered a critical factor in psychiatric disorders. While these different attentional foci are likely generated by the dynamic interplay of multiple cognitive processes, previous studies have generally examined single cognitive dimensions. We developed a new method, cognitive dynamic similarity analysis (C-DSA), to clarify how cognitive processes differ between experimental conditions. In an MR scanner, participants performed a word-processing task in which they focused on either their own emotions or the number of letters associated with a stimulus. To extract cognitive dynamics at the single-trial level, we applied cognitive dynamics estimation, a recently developed method that generates whole-brain activation maps for four cognitive dimensions (emotion processing, selective attention, self-referential thought, and working memory) using a meta-analytic platform. We then performed C-DSA to calculate the difference between internal and external attention for each cognitive dimension. C-DSA revealed significant differences between internal/external attention in all cognitive dimensions, but especially in emotion processing. Moreover, the difference between attention conditions of selective attention was negatively associated with the severity of depression and state-anxiety, but positively associated with trait-anxiety. Our findings suggest that C-DSA applies to both naturalistic and controlled dynamic processes and may be valuable in clinical settings by linking dynamic cognitive mechanisms with issues like ageing and psychiatric disorders.

Predictive influences on memory encoding: investigating oscillations and the N400 event-related potential component
To effectively function in an ever-changing environment, the brain is proposed to make predictions about upcoming information. However, the association between prediction and memory formation and the role of between-subject neural variability in this relationship is unclear. To shed light on the relationship between prediction and memory, the present study reanalysed data from Jano and colleagues (2023). In the original experiment, participants were exposed to naturalistic images in predictable and unpredictable four-item sequences, after which their memory was tested using an old/new paradigm. In the present analysis (N = 46), N400 amplitude and oscillatory power during learning was measured to gauge processes related to prediction error and memory encoding, respectively. This activity was compared with subsequent memory outcomes and individual alpha frequency (IAF) calculated at rest. Linear mixed-effects regressions revealed an alpha power subsequent memory effect that was not related to the amplitude of the N400, suggesting that memory encoding may occur independently of the level of prediction error. Notably, IAF influenced the relationship between theta power, N400 amplitude and subsequent memory, implying that the electrophysiological conditions for successful memory formation differ between individuals. Consequently, the extent to which prediction errors (presumably captured via the N400) drive memory encoding could depend on inter-individual variability in intrinsic neural activity. These findings emphasise the flexible nature of memory, whilst having potential implications for prediction error-driven accounts of learning.

Entorhinal grid-like codes for visual space during memory formation
Eye movements, such as saccades, allow us to gather information about the environment and, in this way, can shape memory. In non-human primates, saccades are associated with the activity of grid cells in the entorhinal cortex. Grid cells are essential for spatial navigation, but whether saccade-based grid-like signals play a role in human memory formation is currently unclear. Here, human participants underwent functional magnetic resonance imaging (fMRI) and continuous eye gaze monitoring while studying scene images. Recognition memory was probed immediately thereafter. Results revealed saccade-based grid-like codes in the left entorhinal cortex while participants studied the scene images, a finding that was replicated with an independent data set reported here. The grid-related effects were time-locked to activation increases in the frontal eye fields. Most importantly, saccade-based grid-like codes were associated with recognition memory, such that grid-like codes were lower the better participants performed in subsequently recognizing the scene images. Collectively, our findings suggest an entorhinal map of visual space that is timed with neural activity in oculomotor regions, supporting memory formation.

A Novel Mouse Model Demonstrates In Vivo Replenishment of Central Nervous System Pericytes After Successful Acute Ablation
Central nervous system (CNS) pericytes play crucial roles in vascular development and blood-brain barrier maturation during prenatal development, as well as in regulating cerebral blood flow in adults. They have also been implicated in the pathogenesis of numerous neurological disorders. However, the behavior of pericytes in the adult brain after injury remains poorly understood, partly due to limitations in existing pericyte ablation models. To investigate pericyte responses following acute ablation, we developed a tamoxifen-inducible pericyte ablation model by crossing PDGFR{beta}-P2A-CreERT2 and Rosa26-DTA176 transgenic mouse lines. Using this model, we studied the effects of different tamoxifen doses and conducted histological examinations 15 and 60 days post-injection to assess both short- and long-term impacts of pericyte ablation. Our results demonstrate that a low dose of tamoxifen effectively ablates CNS pericytes in mice without reducing survival or causing significant systemic side effects, such as weight loss. Additionally, we found that the extent of pericyte depletion varies between the cortex and the spinal cord's gray and white matter regions. Importantly, both pericyte coverage and numbers increased in the weeks following acute ablation, indicating the regenerative capacity of CNS pericytes in vivo. This model offers a valuable tool for future studies on the role of pericytes in neurological disorders, overcoming the limitations of constitutive pericyte ablation models.

Somatostatin interneurons select dorsomedial striatal representations of the initial learning phase
The dorsomedial striatum (DMS) is an associative node involved in the adaptation of ongoing actions to the environmental context and in the initial formation of motor sequences. In early associative or motor learning phases, DMS activity shows a global decrease in neuron firing, eventually giving rise to a select group of active cells, whose number is correlated with animal performance. Unveiling how those representation emerge from DMS circuits is crucial for understanding plasticity mechanisms of early adjustments to learning a task. Here, we hypothesized that inhibitory microcircuits formed by local interneurons are responsible for the genesis of early DMS representation and associated task performance. Despite the low density of somatostatin (SOM)-positive cells, we observed that selective manipulation of SOM cells disrupted reorganization of DMS activity and modulated initial phases of learning in two behavioral contexts. This effect was cell-specific as manipulation of parvalbumin-positive interneurons had no significant effect. Finally, we identified the high plasticity of SOM innervation in the DMS as a key modulator of the SPN excitability and firing activity. Hence, SOM interneurons set the pace of early learning by actively controlling the remapping of DMS network activity.

HEXIM1 is correlated with Alzheimer's disease pathology and regulates immediate early gene dynamics in neurons
Impaired memory formation and recall is a distinguishing feature of Alzheimer's disease, and memory requires de novo gene transcription in neurons. Rapid and robust transcription of many genes is facilitated by their existence in a basal poised state, in which RNA polymerase II (RNAP2) has initiated transcription, but is paused just downstream of the gene promoter. Neuronal depolarization releases the paused RNAP2 to complete the synthesis of messenger RNA (mRNA) transcripts. Paused RNAP2 release is controlled by positive transcription elongation factor b (P-TEFb), which is sequestered into a larger inactive complex containing Hexamethylene bisacetamide inducible protein 1 (HEXIM1) under basal conditions. In this work, we find that neuronal expression of HEXIM1 mRNA is highly correlated with human Alzheimer's disease pathologies. Furthermore, P-TEFb regulation by HEXIM1 has a significant impact on the rapid induction of neuronal gene transcription, particularly in response to repeated neuronal depolarization. These data indicate that HEXIM1/P-TEFb has an important role in inducible gene transcription in neurons, and for setting and resetting the poised state that allows for the robust activation of genes necessary for synaptic plasticity.

Charting Cortical-Layer Specific Area Boundaries Using Gibbs Ringing Attenuated T1w/T2w-FLAIR Myelin MRI
Cortical areas have traditionally been defined by their distinctive layer cyto- and/or myelo- architecture using postmortem histology. Recent studies have delineated many areas by measuring overall cortical myelin content and its spatial gradients using the T1w/T2w ratio MRI in living primates, including humans. While T1w/T2w studies of areal transitions might benefit from using the layer profile of this myelin-related contrast, a significant confound is Gibbs ringing artefact, which produces signal fluctuations resembling cortical layers. Here, we address these issues with a novel approach using cortical layer thickness-adjusted T1w/T2w-FLAIR imaging, which effectively cancels out Gibbs ringing artefacts while enhancing intra-cortical myelin contrast. Whole-brain MRI measures were mapped onto twelve equivolumetric layers, and layer-specific sharp myeloarchitectonic transitions were identified using spatial gradients resulting in a putative 182 area/subarea partition of the macaque cerebral cortex. The myelin maps exhibit unexpectedly high homology with humans suggesting cortical myelin shares the same developmental program across the species. Comparison with histological Gallyas myelin stains explains over 80% of the variance in the laminar T1w/T2w-FLAIR profiles, substantiating the validity of the method. Altogether, our approach provides a novel, noninvasive means for precision mapping layer myeloarchitecture in the primate cerebral cortex, advancing the pioneering work of classical neuroanatomists.